Primary Circumferential Acetabular Labral Reconstruction

Achieving Outcomes Similar to Primary Labral Repair Despite More Challenging Patient Characteristics

John P. Scanaliato,* MD, Daniel L. Christensen,y MD, Catherine Salfiti,z BS, Mackenzie M. Herzog,§ MPH, and Andrew B. Wolff,z|| MD Investigation performed at Washington Orthopaedics and Sports Medicine, Washington, DC, USA

Background: Treatment of acetabular labral tears with moderate or severe intrasubstance damage or segmental defects remains a substantial challenge. Circumferential labral reconstruction with iliotibial band allograft is a relatively new technique that has been proposed to restore stability and eliminate high-stress junction points. Purpose: To compare outcomes between hips treated with primary allograft circumferential labral reconstruction and primary labral repair. Study Design: Cohort study; Level of evidence, 3. Methods: All consecutive hips between 2014 and 2015 that underwent primary reconstruction or primary repair by the senior surgeon were included and compared. Hips that had a prior intra-articular procedure were excluded. Patient-reported outcome (PRO) scores and visual analog scales were completed by patients within 1 week before surgery and between 22 and 26 months postoperatively. PROs included the modified Harris Hip Score, the International Hip Outcome Tool, and the 12-Item Short Form Health Survey for physical health. Pain and satisfaction were assessed with visual analog scales. Crude and inverse probability of treatment weighting comparisons of PROs between groups were performed. Results: A total of 162 hips met the inclusion criteria for this study, including 99 labral repairs and 63 complete labral reconstructions. Patients who underwent labral reconstruction were, on average, older (43.4 vs 29.5 years; P \ .01), had a slightly higher body mass index (24.6 vs 23.0; P \ .01), had hips with a higher To¨ nnis grade (grade 1 or 2: 25% vs 9%; P \ .01), had higher preoperative pain scores (49.9 vs 41.5; P = .01), and had hips with more severe pathology (68% vs 5%; P \ .01) as compared with patients with labral repair hips. Five (5%) labral repair hips and 5 (8%) labral reconstruction hips failed treatment (P = .48). Among hips that did not fail (n = 94 repairs, n = 58 reconstructions), all demonstrated statistically significant improvements in PROs, and there was no statistical difference in PROs between groups after weighting (P . .05). Conclusion: Primary circumferential labral reconstruction is a viable treatment option with promising short-term outcomes for hips that demonstrate moderate or severe labral damage. Despite less favorable preoperative characteristics, labral reconstruction offers similar outcomes when compared with labral repair in hips with less severe pathology. Keywords: femoroacetabular impingement; labral reconstruction; hip arthroscopy

Acetabular labral tears are frequently associated with fem- intrasubstance pathology that are easily treated with repair oroacetabular impingement and can cause significant pain to extensive or complex tears where repair may be difficult and disability.6,31 Treatment of labral tears previously or impossible. included either debridement or labral repair (also described Labral reconstruction has recently been introduced as as labral refixation), with repair typically demonstrating a labral-preserving treatment option for patients with higher function and improved outcomes.8,16,17,22,24 Labral inadequate tissue for repair.33 In the available literature, injury exists on a continuum, from small tears with limited patients with irreparable labral tears or insufficient labral tissue are considered candidates for labral reconstruction.32 Labral reconstruction has led to improvements in symptoms, function, and return to play among elite ath- The American Journal of Sports Medicine letes.2,29 In fact, Domb et al7 demonstrated with a small 1–10 DOI: 10.1177/0363546518775425 cohort that reconstruction was superior to resection, while 19 Ó 2018 The Author(s) Matsuda and Burchette demonstrated that despite more

1 2 Scanaliato et al The American Journal of Sports Medicine significant patient-reported symptoms, reconstruction had The purpose of this study is to compare patient-reported similar outcomes as compared with a repair group. outcomes (PROs) between hips treated with primary allo- Labral reconstruction is hypothesized to restore the nat- graft circumferential labral reconstruction and primary ural biomechanics of the labral suction seal while decreas- labral repair. Secondarily, we assessed the influence of ing pain attributed to resection of the damaged labral tear severity, age, and To¨nnis grade on PROs between tissue—a known pain generator.12 In a cadaveric study, the treatment groups. Our hypothesis was that hips trea- Nepple and colleagues20 and Philippon et al21 found that ted with allograft circumferential labral reconstruction labral reconstruction improves the strength of the labral would have postoperative PROs equivalent to hips treated suction seal, restores stability against distraction to levels with labral repair despite less favorable patient- and hip- similar to those of a native labrum, and restores peak fluid specific preoperative characteristics. pressurization levels to those similar to the intact labrum. Labral reconstruction was also shown to improve the hip joint contact area and contact pressure in an in vitro METHODS model.18 Furthermore, the high level of free nerve ending expression in the labrum suggests that complete resection Patient Population of damaged labral tissue may lead to pain relief.12 Multiple graft options exist for labral reconstruction, All patients scheduled for hip surgery with the senior sur- including autograft iliotibial band, gracilis, semitendino- geon (A.B.W.) are prospectively enrolled in an online out- sus, and ligamentum teres, among others.9,26 However, comes database; these data are collected as standard of there is limited literature on the outcomes of allograft care for all patients. All hips that underwent complete reconstruction.5 Additionally, many labral reconstruction arthroscopic labral reconstruction or arthroscopic labral techniques focus on debridement of the tear back to a stable repair between 2014 and 2015 were selected from the pro- base and creation of a segmental graft that is fixed in this spective outcomes database for inclusion in this study (n = defect. There is concern that the junction points between 210). Patients undergoing a revision procedure were the native labrum and the graft are inherently weak. excluded from this analysis (n = 20). Of the hips that met Shorter segmental reconstruction procedures often posi- the inclusion criteria (n = 190), 4 (2.1%) were excluded tion the junction points between native labrum and graft owing to patient refusal to participate and 24 (12.6%) for in regions of high stress and thus may provide less stability lack of follow-up at postoperative 2 years. The total study or higher risk of breakdown. Furthermore, segmental population comprised 162 hips (85.3% follow-up) among reconstruction techniques may not remove all of the 152 unique patients. This retrospective study was deter- pain-generating tissue and may hinder the ability to per- mined to be exempt from continuing review by the Western form comprehensive pincer-type impingement correction. Institutional Review Board. Circumferential or complete labral reconstruction is per- Indications for hip arthroscopy included hips that had formed by wider excision of the labrum, which eliminates symptoms of sufficient severity and duration to limit desired the labrum-graft interface anteroinferiorly by fixation of activity level and that had failed nonoperative manage- the graft to the acetabulum at the anterior edge of the trans- ment.14 Nonoperative treatment typically spanned a mini- verse acetabular ligament (TAL). Posteriorly, depending on mum of 6 months and consisted of activity modification, the quality of posteroinferior labral tissue, the graft is physical therapy, intra-articular corticosteroid injections, extended either to the TAL or to the 8-o’clock position, and/or nonsteroidal anti-inflammatory medications. Failure where it is fixed to both the labrum and the acetabulum of nonoperative management was defined as continued pain through vertical mattress suture anchor fixation, with the or symptoms despite the aforementioned protocol. Candi- graft abutting the native labrum extra-articularly. dacy for arthroscopic surgery of the hip was further The existing literature has few reports of circumferential assessed by preoperative imaging, which included labral reconstruction with allograft tissue, although White anterior-posterior pelvis, 45° modified Dunn lateral and et al35 demonstrated favorable outcomes at postoperative false-profile view radiographs, and a 3-T magnetic resonance 2 years for patients undergoing labral reconstruction with arthrography.4,13 Treatment of the labrum was determined a similar front-to-back iliotibial band allograft. More intraoperatively: labral tears with mild intrasubstance tear- recently, White et al34 compared 29 patients who underwent ing were repaired, while those with severe intrasubstance primary labral repair on 1 hip and a primary reconstruction damage, labral ossification, or segmental defects were recon- on the other, demonstrating a failure rate of 31% for the structed. In the group with more moderate damage, recon- repair side versus 0% for the reconstruction side. struction was favored for patients .35 years of age and

||Address correspondence to Andrew B. Wolff, MD, Washington Orthopaedics and Sports Medicine, 5215 Loughboro Rd NW, Suite 200, Washington, DC 20016, USA (email: [email protected]). *School of Medicine, Uniformed Services University, Bethesda, Maryland, USA. yDepartment of Orthopaedic Surgery, Walter Reed National Military Medical Center, Bethesda, Maryland, USA. zWashington Orthopaedics and Sports Medicine, Washington, DC, USA. §Professional Research Institute for Sports Medicine, LLC, Chapel Hill, North Carolina, USA. One or more of the authors has declared the following potential conflict of interest or source of funding: A.B.W. is a consultant to Stryker and Arthrex. D.L.C. is a stock owner in Conformis. M.M.H. is owner of the Professional Research Institute for Sports Medicine, LLC, a research consulting company. AJSM Vol. XX, No. X, XXXX Primary CLR vs Labral Repair 3

TABLE 1 Reconstruction vs Repair—Personal Indications and Preferences

Hips that ‘‘always’’ get reconstructed Revision procedures Severe labral damage or missing/ossified labrum Collagen disorders (eg, Ehlers-Danlos) Severe pincer deformities (eg, coxa profunda) Hips in which reconstruction is strongly favored Patients .35 y with moderate labral damage Other considerations to make me favor reconstruction Long history of pain Severe synovitis Mild chondrosis Older female Hips that ‘‘always’’ get repaired Patients \25 y Mild intrasubstance labral damage Other considerations to make me not favor reconstruction Short duration of symptoms (ie, doing a repair or debridement) Coexistent pathology in the hip that appears to be predominate symptom generator Can include patients with far more arthritis than anticipated from preoperative imaging Tumor-like conditions (synovial chondromatosis, pigmented villonodular synovitis, etc)

evidence of some degree of fraying of the substance. Moder- ate labral tears are those with either (1) rim stability and evidence of intrasubstance damage that is more significant than fraying but still \50% of the substance at the site of tearing or (2) an unstable rim with evidence of intrasubstance damage \50% of the substance at the site of tearing. Severe labral tears are complex and multiplane, which include .50% of the labral substance at the site of the tear; stability at the rim may or may not be present. Severe tears are defined principally by the multiplane tearing and significance of the proportion of the substance affected at the tear site. Rim instability is not a prerequisite for a severe tear. The intrarater reliability of tear severity was assessed in a sidearm study, which returned a kappa value of 0.66, indicating moderate reliability. After surgery, the senior surgeon completed a postoperative form capturing intraoperative details, including intraoperative Figure 1. Algorithm for the arthroscopic management of findings, procedures performed, and final diagnoses. acetabular labral tears. Three PROs were used in this study: the modified Har- ris Hip Score (mHHS), the International Hip Outcome Tool patients requiring a larger pincer correction. Repair was (iHOT-12), and the 12-Item Short Form Health Survey for physical health (SF-12 PH). The mHHS is a commonly favored in younger patients with a more predominate cam- 27 type impingement (Table 1, Figure 1). reported PRO for hip joint preservation surgery. The SF-12 PH provides a validated measure of general health with a relatively low responder burden. Finally, the iHOT-12, an abbreviated version of the iHOT-33, captures Data Collection quality of life and changes after intervention in younger, 10,30 Patient characteristics were collected at the initial patient more active patients. Pain was assessed with a visual appointment and entered into an online outcomes database analog scale (VAS) from 0 (no pain) to 100 (worst imagin- as part of our routine standard of care; PROs were col- able pain). Overall satisfaction with surgery was also lected electronically through the online outcomes database assessed with a VAS from 0 (very unsatisfied) to 100 before surgery and at specified postoperative intervals (1 (very satisfied). Failure was defined as need for subsequent and 2 years). To¨nnis grade was determined per preopera- intra-articular hip surgery or total hip arthroplasty (THA). tive radiographs, and labral tear severity was determined arthroscopically by the senior surgeon at the time of sur- Surgical Technique gery. Labral tear severity was classified according to the guidelines established by the Multicenter Arthroscopic All procedures were performed by the senior surgeon. After Study of the Hip (MASH) Study Group. A mild tear is the hip joint was accessed arthroscopically, the extent of described as a labrum with a stable rim configuration but the labral tear was assessed by visual inspection and 4 Scanaliato et al The American Journal of Sports Medicine

Figure 2. Pre- and postoperative labral reconstruction arthroscopic photographs. Top row: Before labral treatment. (A) A superior/posterior view demonstrating severe intrasub- Figure 3. Postreconstruction arthroscopic view of the labral stance tearing deemed irreparable. (B) A superior/anterior suction seal. Restoration of the labral suction seal after trac- view demonstrating more moderate tearing. Overall classifi- tion released. cation for this degree of labral pathology is severe. Bottom row: After labral treatment. (C, D) Views of the same patient inferior anchor, adjacent to the TAL. The remaining sutures after circumferential labral reconstruction. (typically 9-12 in total) were then passed around and/or through the graft. Posteriorly, any remaining graft was probing, with careful observation for pathology identified sutured in a vertical mattress to the remaining posteroinfe- on preoperative imaging. If the tissue was deemed repara- rior labrum at approximately the 8-o’clock position with 1 ble, the acetabular rim was decorticated to a bed of bleed- or 2 suture anchors, with the remaining graft abutting the ing bone, or a pincer correction was performed if indicated, native labrum extra-articularly. This technique obviates the with careful attention to protect the labrum from further need to have a precise measurement of the defect, which damage. Suture anchors were placed into the acetabulum has proven challenging (Figure 2). as close to the rim as possible, and the sutures from the After labral treatment, the hip was taken out of traction anchors were passed in a vertical mattress or simple loop and brought through range of motion to ensure a complete configuration depending on the labral anatomy. A simple restoration of the suction seal, as demonstrated in Figure loop was used for more cylindrical and/or smaller labra, 3. The capsule was either repaired or plicated in all whereas vertical mattress sutures were preferred for patients in this study. Postoperatively, passive motion larger and/or meniscoid labra. began immediately with a continuous passive motion For the labral reconstruction technique, damaged native machine or with a resistance-less stationary bike. Patients labral tissue was excised between the TAL anteroinferiorly in both groups were instructed to use protected weight- and the 8-o’clock position posteroinferiorly. If there was lab- bearing with crutches until their gait normalized, typically ral damage or ossification that extended farther than the 2 to 4 weeks postoperatively. 8-o’clock position, the resection and subsequent reconstruction were taken down to the level of the TAL posteroinfer- Statistical Analysis iorly. The rim of the acetabulum was decorticated throughout the length of the proposed reconstruction. Bony All statistical analyses were performed with SAS (v 9.4; impingement was corrected as indicated. On the back table, SAS Institute), and a P value \.05 was considered statisti- a fascia lata allograft was tubularized to make a 6-mm diam- cally significant. Comparisons of study population charac- eter graft with a length between 85 and 110 mm. Length was teristics and concomitant procedures between the labral determined by experience of the surgeon with the goal of treatment groups were performed with 2-tailed Student t never having a graft that is too short based on acetabular tests for continuous data and Fisher exact tests for categor- size and length of the defect. In this study, all grafts started ical data. Pre- to postoperative PROs were compared anteriorly at the TAL and ended by overlapping with the within groups via paired Student t tests. To evaluate the native labrum posteroinferiorly (at or below 8 o’clock) or at average difference in failure rate and postoperative PROs the TAL posteroinferiorly. No grafts in this study were too between treatment groups, log-binomial or linear regressions short; thus, none left a segmental defect. Suture anchors were used to generate point estimates (beta) and 95% CIs. were placed into the acetabular rim, and the graft was passed Inherent differences exist between the hips that were through a cannula down the suture from the most anterior- selected to undergo labral repair and labral reconstruction AJSM Vol. XX, No. X, XXXX Primary CLR vs Labral Repair 5

TABLE 2 Study Population Characteristics Stratified by Labral Treatmenta

Repair (n = 99) Reconstruction (n = 63) P Value

Age, y 29.5 6 11.0 43.4 6 10.7 \.01 Male 42 (42.4) 26 (41.3) .89 Body mass index 23.0 6 3.5 24.6 6 3.8 .01 Severity Minimal 10 (10.1) 0 (0.0) \.01 Moderate 84 (84.8) 20 (31.7) Severe 5 (5.1) 43 (68.3) To¨nnis grade 0 90 (90.9) 47 (74.6) \.01 1 9 (9.1) 8 (12.7) 2 0 (0.0) 8 (12.7) Preoperative measure mHHS 63.4 6 17.1 60.2 6 15.5 .24 iHOT-12 39.3 6 18.0 37.8 6 19.7 .62 SF-12 PH 37.4 6 8.3 37.6 6 9.4 .89 VAS pain 41.5 6 19.9 49.9 6 21.7 .01

aData are provided as No. of hips (%) or mean 6 SD. iHOT-12, International Hip Outcome Tool; mHHS, modified Harris Hip Score; SF-12 PH, 12-Item Short Form Health Survey for physical health; VAS, visual analog scale. such that the labral reconstruction hips typically had lower RESULTS preoperative subjective scores, more pain, and more intra- articular damage. Subsequently, crude comparisons of out- A total of 162 hips met the inclusion criteria for this study, comes are limited. To adjust for some of the differences in including 99 labral repairs and 63 complete labral recon- baseline characteristics, inverse probability of treatment structions with iliotibial band allograft. Hips that under- weighting (IPTW) was used to compare PROs between went labral reconstruction were on average older (43.4 vs treatment groups.15 Each hip was assigned a probability 29.5 years; P \ .01); had a higher BMI (24.6 vs 23.0; P \ of undergoing labral reconstruction (ie, a propensity score) .01), a higher To¨nnis grade (grade 1 or 2: 25% vs 9%; P \ based on the following preoperative characteristics: age at .01), and higher preoperative pain scores (49.9 vs 41.5; the time of surgery, body mass index (BMI), To¨nnis grade, P = .01); and were more likely to be rated as having severe preoperative pain score, and preoperative mHHS. Propen- labral pathology, defined principally as a multiplane tear sity scores were first described by Rosenbaum and Rubin23 with or without rim instability (68% vs 5%; P \ .01) as in 1983 as a method to address confounding between treat- compared with labral repair hips (Table 2). Labral recon- ment groups, and since their development, propensity struction hips also had more concomitant procedures per- scores have been widely used in nonexperimental studies formed, on average, than labral repair hips (Table 3). comparing outcomes after medical interventions. The pro- Among the 162 hips with 2-year follow-up data collected pensity score for each hip was calculated with a logistic between 22 and 26 months, the mean follow-up time was regression model, which calculated a probability of the 24.0 months (SD = 1.9 months). Five (5%) labral repair exposure (labral reconstruction) given the preoperative hips and 5 (8%) labral reconstruction hips failed treatment characteristics (ie, age at the time of surgery, BMI, To¨nnis (P = .48). Of those, 4 labral repair hips required revision grade, preoperative pain score, and preoperative mHHS). arthroscopic treatment and 1 required THA, while 3 labral Age and BMI were treated as categorical variables because reconstruction hips required revision arthroscopic treat- of the limited sample size. With the calculated probability ment and 2 required THA. The hips that failed treatment of undergoing a labral reconstruction, hips were weighted were not included in the analysis of PROs. to create groups that were most similar preoperatively by Among hips that did not fail treatment (n = 94 repairs, applying a weight equal to the inverse of the propensity n = 58 reconstructions), all patients demonstrated statisti- score (ie, IPTW). Hips that underwent labral reconstruc- cally significant improvements between pre- and postoper- tion were assigned a weight equal to 1 / propensity score, ative PRO scores (Table 4). Crude comparison of labral and hips that underwent labral repair were assigned repair hips with labral reconstruction hips suggests that a weight equal to 1 / (1 2 propensity score). The weights repair hips have higher postoperative PRO scores and sat- were stabilized according to the methods described by isfaction and lower pain scores (Table 5), yet these differen- Sturmer et al28 in 2010, and hips without similar compara- ces diminished with IPTW after adjusting for differences in tors between groups were trimmed out of the analytic pop- preoperative characteristics. ulation, resulting in a restricted but more comparable Results from IPTW yielded an adjusted population of study population. Differences in PROs between treatment 128 hips, with labral repair and labral reconstruction groups were then assessed with adjusted betas and 95% hips that were preoperatively similar in terms of age, CIs generated with IPTW linear regressions. BMI, To¨nnis grade, and preoperative mHHS and pain 6 Scanaliato et al The American Journal of Sports Medicine

TABLE 3 Concomitant Procedures Stratified by Labral Treatmenta

Procedure Repair (n = 99) Reconstruction (n = 63) P Value

Osteoplasty Cam 81 (81.8) 61 (96.8) \.01 Pincer 28 (28.3) 44 (69.8) \.01 Subspine 2 (2.0) 5 (7.9) .07 Fovea 0 (0.0) 2 (3.2) .07 Chondroplasty Acetabular 54 (54.5) 56 (88.9) \.01 Femoral head 7 (7.1) 9 (14.3) .13 Unspecified 2 (2.0) 2 (3.2) .64 Acetabular microfracture 0 (0.0) 1 (1.6) .21 Excision of acetabular rim fracture 3 (3.0) 6 (9.5) .08 Labral debridement 3 (3.0) NA NA Debridement of intralabral calcifications 1 (1.0) NA NA Ligamentum teres debridement 23 (23.2) 19 (30.2) .33 Synovectomy 39 (39.4) 38 (60.3) \.01 Capsular plication 27 (27.3) 11 (17.5) .15 Loose body removal 5 (5.1) 6 (9.5) .27 Drilling of subchondral cyst 1 (1.0) 1 (1.6) .75 Synovial biopsy 2 (2.0) 0 (0.0) .26 Excision Ganglion cyst 1 (1.0) 0 (0.0) .42 Heterotopic ossification 1 (1.0) 0 (0.0) .42 Trochanteric bursectomy 2 (2.0) 10 (15.9) \.01 Repair of gluteus medius and/or minimus tear 0 (0.0) 3 (4.8) .03 Lesser trochanter osteoplasty 2 (2.0) 1 (1.6) .84

aNA, not applicable.

TABLE 4 Change in Patient-Reported Outcome Scores Stratified by Treatment Groupa

Outcome Measure Preoperative Postoperative Differenceb 95% CI

Repair (n = 94) mHHS 63.4 6 17.1 88.0 6 15.1 24.2 20.3 to 28.1 iHOT-12 39.3 6 18.0 71.2 6 23.2 31.7 26.9 to 36.5 SF-12 PH 37.4 6 8.3 50.0 6 9.0 12.7 10.6 to 14.7 VAS pain 41.5 6 19.9 14.1 6 17.1 –27.7 232.2 to 223.3 Reconstruction (n = 58) mHHS 60.2 6 15.5 80.7 6 16.4 20.4 15.6 to 25.1 iHOT-12 37.8 6 19.7 65.8 6 26.2 27.8 19.7 to 35.8 SF-12 PH 37.6 6 9.4 47.1 6 10.1 9.3 6.6 to 12.0 VAS pain 49.9 6 21.7 23.6 6 22.5 –25.6 233.1 to 218.1

aScores are given as mean 6 SD. iHOT-12, International Hip Outcome Tool; mHHS, modified Harris Hip Score; SF-12 PH, 12-Item Short Form Health Survey for physical health; VAS, visual analog scale. bThe difference is the mean postoperative score minus the mean preoperative score. All P values for differences, P \ .01. scores (Table 6). When outcomes were compared between difference, and the point estimate for postoperative pain hips that underwent labral reconstruction and labral moved from an average 10-point difference in pain score repair with this weighted population, there was no statis- between groups to an average 4-point difference. tical difference in outcome scores between treatment groups (Table 7). In addition, there was a trend toward improvement on all measures except for patient satisfaction for the labral reconstruction group, although these DISCUSSION changes were not statistically significant. For example, the point estimate for mHHS moved from an average 26- For primary arthroscopic treatment of labral pathology of the point difference between groups to an average 23-point hip, labral reconstruction can provide similar outcomes to AJSM Vol. XX, No. X, XXXX Primary CLR vs Labral Repair 7

TABLE 5 Postoperative Outcome Measures Stratified by Labral Treatment (n = 152)a

Repair (n = 94) Reconstruction (n = 58) Differenceb 95% CI P Value mHHS 88.0 6 15.1 80.7 6 16.4 –7.3 212.4 to 22.1 \.01 iHOT-12 71.2 6 23.2 65.8 6 26.2 –5.5 –13.4 to 2.5 .18 SF-12 PH 50.0 6 9.0 47.1 6 10.1 –2.9 –6.0 to 0.2 .07 VAS pain 14.1 6 17.1 23.6 6 22.5 9.5 3.2 to 15.8 \.01 VAS satisfaction 81.9 6 24.2 77.1 6 27.0 –4.7 –13 to 3.6 .27

TABLE 6 Restricted Study Population Characteristics Stratified by Labral Treatment: Pre– and Post–Propensity Score Weighting (n = 128)a

Restricted Study Population: Raw Values IPTW Study Population: Adjusted Values

Preoperative Characteristic Repair (n = 80) Reconstruction (n = 48) P Value Repair Reconstruction P Value

Age, y \.001 .40 \30 41 (51.2) 5 (10.4) 28.3 (34.5) 32.5 (52.3) 30-39.9 23 (28.7) 13 (27.1) 20.7 (25.2) 12.5 (20.1) 40-49.9 11 (13.8) 21 (43.8) 24.7 (30.1) 12.1 (19.4) 50 5 (6.2) 9 (18.8) 8.4 (10.2) 5.1 (8.2) Body mass index .71 .40 \20 5 (6.2) 4 (8.3) 6.7 (8.1) 2.9 (4.7) 20-24.9 58 (72.5) 30 (62.5) 58.4 (71.2) 52.1 (83.8) 25-29.9 12 (15.0) 10 (20.8) 11.8 (14.4) 4.8 (7.8) 30 5 (6.2) 4 (8.3) 5.2 (6.3) 2.3 (3.7) To¨nnis grade .0 7 (8.8) 7 (14.6) .47 12.1 (14.7) 5.8 (9.3) .49 Preoperative mHHS 63.7 6 17.9 61.9 6 16.2 .57 63.5 6 17.4 65.1 6 12.0 .54 VAS pain 45.3 6 19.4 45.6 6 20.3 .93 46.0 6 18.2 40.3 6 17.1 .15

aValues are presented as n (%) or mean 6 SD. IPTW, inverse probability of treatment weighted; mHHS, modified Harris Hip Score; VAS, visual analog scale.

TABLE 7 Adjusted Postoperative Outcome Measures Stratified by Labral Treatment in the Restricted Population (n = 128)a

Unadjusted Adjusted With IPTW

Differenceb 95% CI P Value Differenceb 95% CI P Value mHHS –6.4 212.1 to 20.7 .03 –3.0 –8.4 to 2.4 .28 iHOT-12 –7.3 –15.9 to 1.4 .10 –2.8 –10.1 to 4.6 .46 SF-12 PH –3.0 –6.5 to 0.5 .10 0.4 –2.7 to 3.4 .81 VAS pain 9.8 2.8 to 16.9 .01 4.1 –2 to 10.2 .19 Satisfaction –3.6 –12.5 to 5.3 .43 –4.9 –12.5 to 2.7 .21

aiHOT-12, International Hip Outcome Tool; IPTW, inverse probability of treatment weighted; mHHS, modified Harris Hip Score; SF-12 PH, 12-Item Short Form Health Survey for physical health; VAS, visual analog scale. bThe difference is the mean reconstruction score minus the mean repair score; negative values indicate that the reconstruction group had lower scores, on average, than the repair group. labral repair. These results are especially promising when significantly older, with a higher BMI, more severe labral the characteristics and pathology of the two groups are com- tears, and higher To¨nnis grades. These hips also had lower pared. The hips in the labral reconstruction group were preoperative mHHS scores and higher preoperative pain 8 Scanaliato et al The American Journal of Sports Medicine

VAS scores. As highlighted in Table 3, reconstruction hips remains unknown. However, given the similarities between generally had more concomitant procedures performed, labral and meniscal tissue and the poor outcomes associated which suggests a worse preoperative intra-articular problem with repair of severely damaged meniscal tissue, it is logical that necessitated a more complicated surgical procedure. At that less favorable results could be expected with more 2-year follow-up, there was no statistically significant differ- severely damaged labral tissue. ence in mHHS, iHOT-12, SF-12 PH, or pain VAS between Labral reconstruction is a technically difficult procedure the treatment groups, although repair hips tended to have that has the potential to offer patients biomechanical sta- slightly higher postoperative PROs and lower postoperative bility and pain relief in settings where repair results may pain.Therewasalsonodifferenceinfailureratesbetween not be favorable. In our practice, labral repair is preferred the groups. In addition, there was no statistically significant for patients with an unstable or detached labral base with difference in satisfaction at 2-year follow-up between the only mild or moderate intrasubstance tearing. Labral groups. For all-comers, there appears to be no difference in reconstruction is preferred for patients with segmental lab- PROs between labral reconstruction and labral repair hips, ral defects, severe intrasubstance damage, labral ossifica- suggesting that primary labral reconstruction may be an tion, or severe pincer deformities; for older patients for important treatment consideration for hips with significant whom there is concern regarding the healing of repaired labral pathology—particularly in the setting of other, less tissue; or in cases when excessive synovitis concerning favorable characteristics, including advanced age, increased for chronic pathology is identified. A labral repair may BMI, and increased chondrosis. restore anatomic function but fail to remove pain-generating In our clinical experience, it has been a constant challenge tissue. In contrast, labral reconstruction, while far more tech- predicting which hips will do well with labral repair alone nically demanding, not only restores anatomic function but versus those that may require a complete resection of native removes pain-generating tissue. This assertion is supported labral tissue with subsequent circumferential reconstruction. by our data, which demonstrated a statistically significant As our surgical technique continues to improve and our improvement in pain-level and hip-specific outcomes, sug- understanding of hip and labral pathology grows, so does gesting that the complete removal of pain-generating tissue our clinical judgment regarding patient selection for hip has a marked effect on overall hip health. It is also important arthroscopy. Chandrasekaran et al3 found that patients to consider the direct costs associated with each procedure. with To¨nnis grade 2 osteoarthritis of the hip who were under- Labral repair necessitates the use of multiple arthroscopic going hip arthroscopy had significantly higher rates of failure anchors, and labral reconstruction typically requires more with eventual conversion to THA when compared with anchors and allograft tissue. These costs, while not insignifi- patients with To¨nnis grade 0 and 1 hips. In addition, cant, are marginal when compared with the cost of a revision a recently published review of the literature for outcomes procedure and the personal and societal costs associated with after hip arthroscopy in patients aged 40 years revealed failed surgery; surgeons should accordingly consider these the overall reoperation rate for all-comers to be 20.8%, with costs when evaluating outcomes and the survivability of risk factors for revision including increased patient age, these procedures in short- and long-term settings. a higher severity of arthrosis, and lower preoperative out- Two of the main technical decisions for labral reconstruc- come scores.11 tion are length of reconstruction and graft choice. All labral These findings suggest that a subset of hips has severe reconstructions performed in this study utilized a fascia lata intra-articular injury in a traditionally difficult-to-treat allograft, which, in our clinical experience, offers the senior patient group that may not find success with labral repair. surgeon the highest level of consistency with regard to final We believe that primary complete labral reconstruction graft shape, size, and quality. Although the literature on should be considered for these patients with a known pro- anterior cruciate ligament (ACL) reconstruction suggests pensity for suboptimal outcomes with labral repair, which, a higher failure rate for allograft reconstructions versus in our clinical experience, are older patients with more autograft, the ACL is subject to a different quantity and severe labral damage. Primary labral reconstruction, in quality of forces than the acetabular labrum; as such, the a specific patient population, has the ability to offer simi- applicability of the allograft ACL literature to labral recon- larly good outcomes when compared with younger patients struction is questionable.1 In terms of graft length, both the undergoing labral repair for milder labral injury. Even senior surgeon and the existing literature have indicated more promising is that at postoperative 2 years, none of that longer grafts tend to lead to more favorable out- the To¨nnis grade 1 and 2 hips included in this study comes.32,34-36 This relationship likely exists for several rea- exhibited signs of failure suggestive of the need for THA. sons: the circumferential labrum graft has either no The acetabular labrum is a fibrocartilaginous structure junction points or 1 far posteroinferiorly in an area of low similar in composition to the meniscus of the human stress, and, as the labrum is a known pain generator, resect- knee.25 It is well described in the knee literature that there ing more damaged labral tissue should result in decreased exists a continuum of damage to meniscal tissue along pain for the patient. Finally, complete labral reconstruction which a patient moves from a candidate for a meniscal reduces the potential for undercorrection of a pincer defor- repair to one for debridement only.37 This is also true of lab- mity or undertreatment of an ossified labrum or os acetabuli ral tears, with clinical presentation ranging from asymp- owing to concerns about labral preservation or a lack of tomatic to debilitating pain. Despite the success achieved exposure of the acetabular rim. with repair of labral tissue with more minor tears, the appli- It is important to note that direct comparison of surgical cability of labral repair as a treatment for severe tears techniques within the study population with severe labral AJSM Vol. XX, No. X, XXXX Primary CLR vs Labral Repair 9 pathology is limited in this study, given that treatment was cohort study. Given the technical demands of circumferen- influenced by preoperative characteristics. In particular, tial labral reconstruction, the results of this study may not the hips that underwent primary labral reconstruction ver- be generalizable to other patient populations. sus labral repair had preoperative characteristics that may predispose them to poorer outcomes. However, the advanced analytic method of IPTW was used to adjust for some of these CONCLUSION differences in preoperative characteristics to facilitate comparison. This analytic technique weights the labral repair Primary circumferential labral reconstruction is a viable and labral reconstruction groups toward a ‘‘middle ground’’ treatment option with promising short-term outcomes for of preoperative characteristics, which helps to compare the hips that demonstrate moderate or severe labral damage. 2 procedures among a group of hips with more moderate pre- Despite less favorable preoperative characteristics, labral operative patient characteristics. The results suggest that reconstruction offers similar outcomes when compared when some of the preoperative differences are accounted with labral repair in hips with less severe pathology. for, reconstruction outcomes move further toward equiva- lence or potentially improved outcomes over repair. Note that not all predictors of treatment choice were able to be ACKNOWLEDGMENT measured and accounted for, as treatment choice is largely a subjective decision by the surgeon and patient. Optimiza- The authors acknowledge the hard work of all members of tion of treatment prediction, including better measurement the Multicenter Arthroscopic Study of the Hip Study of preoperatively identified pathology and subjective sympto- Group, whose contribution of the data collection infrastruc- mology, may have improved our ability to compare treatment ture utilized in this project was instrumental. options. Nonetheless, we believe it is meaningful that the primary labral reconstruction groups, with greater known predictors for poorer outcomes, had outcomes equivalent to those of the primary labral repair group, which had less REFERENCES severe pre- and intraoperative pathology. 1. Bottoni CR, Smith EL, Shaha J, et al. Autograft versus allograft anterior cruciate ligament reconstruction: a prospective, randomized clin- Limitations ical study with a minimum 10-year follow-up. Am J Sports Med. 2015;43(10):2501-2509. This study is not without limitations. As in nearly all con- 2. Boykin RE, Patterson D, Briggs KK, Dee A, Philippon MJ. Results of temporary literature regarding arthroscopic hip surgery, arthroscopic labral reconstruction of the hip in elite athletes. Am J Sports Med. 2013;41(10):2296-2301. there are often multiple concomitant procedures performed 3. Chandrasekaran S, Darwish N, Gui C, Lodhia P, Suarez-Ahedo C, at the time of surgery. Attributing outcomes to the labral Domb BG. Outcomes of hip arthroscopy in patients with To¨ nnis treatment can present a challenge; however, indications grade-2 osteoarthritis at a mean 2-year follow-up: evaluation using for treatment of concomitant pathology do not vary a matched-pair analysis with To¨ nnis grade-0 and grade-1 cohorts. between labral treatment groups, which should reduce J Bone Joint Surg Am. 2016;98(12):973-982. bias in the comparison of outcomes. As mentioned, it is 4. Clohisy JC, Carlisle JC, Beaule PE, et al. A systematic approach to the plain radiographic evaluation of the young adult hip. J Bone Joint also likely that confounding between treatment groups Surg Am. 2008;90(suppl 4):47-66. remains, despite attempts to adjust for some known con- 5. Costa Rocha P, Klingenstein G, Ganz R, Kelly BT, Leunig M. Circum- founding factors with IPTW. While it is possible that a ran- ferential reconstruction of severe acetabular labral damage using domized controlled trial comparing the outcomes of labral hamstring allograft: surgical technique and case series. Hip Int. repair with labral reconstruction could eliminate much of 2013;23(suppl 9):S42-S53. this confounding, the recent evolution in hip arthroscopy 6. Dolan MM, Heyworth BE, Bedi A, Duke G, Kelly BT. CT reveals a high and continuous generation of evidence regarding proce- incidence of osseous abnormalities in hips with labral tears. Clin Orthop Relat Res. 2011;469(3):831-838. dure outcomes in specific settings make it difficult for 7. Domb BG, El Bitar YF, Stake CE, Trenga AP, Jackson TJ, Lindner D. many surgeons, including the senior author (A.B.W.), to Arthroscopic labral reconstruction is superior to segmental resection justify randomization. As randomized controlled trials for irreparable labral tears in the hip: a matched-pair controlled study are costly and require a long study period, nonexperimen- with minimum 2-year follow-up. Am J Sports Med. 2014;42(1):122-130. tal studies like the present one can generate important 8. Espinosa N, Beck M, Rothenfluh DA, Ganz R, Leunig M. Treatment of real-world evidence to answer such questions if differences femoro-acetabular impingement: preliminary results of labral refixation. Surgical technique. J Bone Joint Surg Am. 2007;89(suppl 2, pt in treatment groups are addressed through, for example, 1):36-53. implementation of IPTW. It is also important to acknowl- 9. Geyer MR, Philippon MJ, Fagrelius TS, Briggs KK. Acetabular labral edge that this is a study of early outcomes. In particular, reconstruction with an iliotibial band autograft: outcome and survi- our mean follow-up of 2 years may not allow for enough vorship analysis at minimum 3-year follow-up. Am J Sports Med. time to fully capture revisions or failures. Furthermore, 2013;41(8):1750-1756. radiographic measurements of femoroacetabular impinge- 10. Griffin DR, Parsons N, Mohtadi NG, Safran MR. A short version of the International Hip Outcome Tool (iHOT-12) for use in routine clinical ment, such as alpha angle and lateral center-edge angle, practice. Arthroscopy. 2012;28(5):611-616. were not included in our analysis. Finally, a single surgeon 11. Griffin DW, Kinnard MJ, Formby PM, McCabe MP, Anderson TD. with extensive hip arthroscopy experience in a practice Outcomes of hip arthroscopy in the older adult: a systematic review dedicated solely to hip arthroscopy treated all hips in this of the literature. Am J Sports Med. 2017;45(8):1928-1936. 10 Scanaliato et al The American Journal of Sports Medicine

12. Haversath M, Hanke J, Landgraeber S, et al. The distribution of noci- 24. Schilders E, Dimitrakopoulou A, Bismil Q, Marchant P, Cooke C. ceptive innervation in the painful hip: a histological investigation. Arthroscopic treatment of labral tears in femoroacetabular impinge- Bone Joint J. 2013;95(6):770-776. ment: a comparative study of refixation and resection with a minimum 13. Jesse MK, Petersen B, Strickland C, Mei-Dan O. Normal anatomy two-year follow-up. J Bone Joint Surg Br. 2011;93(8):1027-1032. and imaging of the hip: emphasis on impingement assessment. 25. Seldes RM, Tan V, Hunt J, Katz M, Winiarsky R, Fitzgerald RH Jr. Semin Musculoskelet Radiol. 2013;17(3):229-247. Anatomy, histologic features, and vascularity of the adult acetabular 14. Kivlan BR, Nho SJ, Christoforetti JJ, et al. Multicenter outcomes after labrum. Clin Orthop Relat Res. 2001;382:232-240. hip arthroscopy: epidemiology (MASH Study Group). What are we 26. Sierra RJ, Trousdale RT. Labral reconstruction using the ligamentum seeing in the office, and who are we choosing to treat? Am J Orthop teres capitis: report of a new technique. Clin Orthop Relat Res. (Belle Mead NJ). 2017;46(1):35-41. 2009;467(3):753-759. 15. Kurth T, Walker AM, Glynn RJ, et al. Results of multivariable logistic 27. Smith MV, Klein SE, Clohisy JC, Baca GR, Brophy RH, Wright RW. regression, propensity matching, propensity adjustment, and Lower extremity-specific measures of disability and outcomes in propensity-based weighting under conditions of nonuniform effect. orthopaedic surgery. J Bone Joint Surg Am. 2012;94(5):468-477. Am J Epidemiol. 2006;163(3):262-270. 28. Sturmer T, Rothman KJ, Avorn J, Glynn RJ. Treatment effects in the 16. Larson CM, Giveans MR. Arthroscopic debridement versus refixation presence of unmeasured confounding: dealing with observations in of the acetabular labrum associated with femoroacetabular impinge- the tails of the propensity score distribution—a simulation study. ment. Arthroscopy. 2009;25(4):369-376. Am J Epidemiol. 2010;172(7):843-854. 17. Larson CM, Giveans MR, Stone RM. Arthroscopic debridement versus 29. Walker JA, Pagnotto M, Trousdale RT, Sierra RJ. Preliminary pain and refixation of the acetabular labrum associated with femoroacetabular function after labral reconstruction during femoroacetabular impinge- impingement: mean 3.5-year follow-up. Am J Sports Med. ment surgery. Clin Orthop Relat Res. 2012;470(12):3414-3420. 2012;40(5):1015-1021. 30. Ware J Jr, Kosinski M, Keller SD. A 12-Item Short-Form Health Sur- 18. Lee S, Wuerz TH, Shewman E, et al. Labral reconstruction with iliotib- vey: construction of scales and preliminary tests of reliability and val- ial band autografts and semitendinosus allografts improves hip joint idity. Med Care. 1996;34(3):220-233. contact area and contact pressure: an in vitro analysis. Am J Sports 31. Wenger DE, Kendell KR, Miner MR, Trousdale RT. Acetabular labral Med. 2015;43(1):98-104. tears rarely occur in the absence of bony abnormalities. Clin Orthop 19. Matsuda DK, Burchette RJ. Arthroscopic hip labral reconstruction Relat Res. 2004;426:145-150. with a gracilis autograft versus labral refixation: 2-year minimum out- 32. White BJ, Herzog MM. Labral reconstruction: when to perform and comes. Am J Sports Med. 2013;41(5):980-987. how. Front Surg. 2015;2:27. 20. Nepple JJ, Philippon MJ, Campbell KJ, et al. The hip fluid seal—part 33. White BJ, Herzog MM. Arthroscopic labral reconstruction of the hip II: the effect of an acetabular labral tear, repair, resection, and recon- using iliotibial band allograft and front-to-back fixation technique. struction on hip stability to distraction. Knee Surg Sports Traumatol Arthrosc Tech. 2016;5(1):e89-e97. Arthrosc. 2014;22(4):730-736. 34. White BJ, Patterson J, Herzog MM. Bilateral hip arthroscopy: direct 21. Philippon MJ, Nepple JJ, Campbell KJ, et al. The hip fluid seal—part comparison of primary acetabular labral repair and primary acetabu- I: the effect of an acetabular labral tear, repair, resection, and recon- lar labral reconstruction. Arthroscopy. 2018;34(2):433-440. struction on hip fluid pressurization. Knee Surg Sports Traumatol 35. White BJ, Stapleford AB, Hawkes TK, Finger MJ, Herzog MM. Allo- Arthrosc. 2014;22(4):722-729. graft use in arthroscopic labral reconstruction of the hip with front- 22. Philippon MJ, Weiss DR, Kuppersmith DA, Briggs KK, Hay CJ. to-back fixation technique: minimum 2-year follow-up. Arthroscopy. Arthroscopic labral repair and treatment of femoroacetabular 2016;32(1):26-32. impingement in professional hockey players. Am J Sports Med. 36. Wolff AB, Grossman J. Management of the acetabular labrum. Clin 2010;38(1):99-104. Sports Med. 2016;35(3):345-360. 23. Rosenbaum P, Rubin D. The central role of the propensity score in 37. Yoon KH, Park KH. Meniscal repair. Knee Surg Relat Res. observational studies for causal effects. Biometrika. 1983;70(1):41-55. 2014;26(2):68-76.

For reprints and permission queries, please visit SAGE’s Web site at http://www.sagepub.com/journalsPermissions.nav.